Defect Detection for Gear System of Station Wagon by Extensive Empirical Wavelet Decomposition and Deep Extreme Learning Machine

Abstract

Abstract Gear system is the one of the most important components of station wagon, so it is very important to study the defect detection method for gear systemof station wagon. Defect detection for gear system of station wagon by extensive empirical wavelet decomposition and deep extreme learning machine is proposed in this paper. As the features the time-frequency image based on extensive empirical wavelet decomposition (EEWT) are clearer than those of empirical wavelet decomposition (EWT), EEWT is used to decompose the vibration signals of gear system of station wagon, and create the time-frequency images of the vibration signals of gear system of station wagon. Deep extreme learning machine (DELM) is formed by stacking multi-layer extreme learning auto-encoders, so it can extract higher-level features and has higher classification and recognition accuracy than traditional ELM, thus, DELM is used to defect detection for gear system of station wagon. The experimental results demonstrates that the defect detectionaccuracy of EEWT-DELM is higher than EWT-DELM,EWT-ELM, and traditional ELM, and EEWT-DELM is suitable for defect detection of gear system of station wagon.